Workshop Phonological Typology of Syllable and Word Languages in Theory and Practice, Frias, Freiburg i.B., 29-31 March 2010.

- 1-

Opposite typological developments in French and Dutch Roland Noske

Université Charles-de-Gaulle Lille 3 & UMR 8163 ‘STL’ roland.noske@univ-lille3.fr, http://perso.univ-lille3.fr/~rnoske

Table 1: prototypical properties of syllable (syllable counting) versus word (stress counting) languages1

nr. criterion syllable languages → syllable counting syllable as basic prosodic unit (foot length variable)

word / accent languages → stress counting phonological word as basic prosodic unit (syllable length variable)

1 syllable structure

CV syllables (rarely closed syllables); all syllables equally long

variable syllables type of different complexity, dependent on the stress position; often differences between medial and peripheral syllables

2 syllable boundaries

well defined, constant syllable boundaries

ill-defined, variable, speech-rate dependent syllable boundaries

3 sonority hierarchy

sonority hierarchy is obeyed, i.e. maximal sonority difference between C and V

sonority hierarchy is less obeyed, e.g. voicing of intervocalic plosives, assimilations (word internally).

4 geminates geminates possible geminate reduction, except in places where they are morphologically relevant, e.g. in internal compound boundaries e.g. German Schifffahrt [f]

5 stress effects no / few differences in structure of stressed vs. unstressed syllables

stressed syllables are heavy, unstressed syllables are light

6 stress assignment

mostly syllable based; absence of fixed word stress possible

stress assignment (often complex) is morphologically / lexically / semantically determined

7 tonality can be present, also on unstressed syllables

if present (which is rarely the case), then only on stressed syllables

8 phonotactics regular, stable phonotactics, no positionally determined allophones

word boundary (delimitative) signals positionally determined allophone (initial, medial, final) phonotactic restrictions

9 vocalism little discrepancy between strongly and weakly stressed syllables, relatively equal tenseness.

strong discrepancy between strongly and weakly stressed vowels (German, Danish, English). Heavy stress: often difference in length, centralizations (reductions)

10 vowel harmony / umlaut

possible rare

11 vowel reduction and deletion

because of reasons of syllable optimization

because of stress

1 This table is has been adapted from Nübling & Schrambke (2004: 284-285).

Roland Noske

- 2 -

12 epenthesis for reason of syllable optimization comp. epenthetic e in Luxemb. Arem, hëllefen, Vollek, intrusive n in Allemanic

if there is, then in order to let stand out morphemic structures like in German eigen-t-lich, namen-t-lich, etc, bonding phoneme s in German and Dutch

13 liaison yes (across morpheme boundaries)

no (border signals / junctures, e.g. glottal stop)

14 sandhi external internal 15 consequences

for morphology

morphs that promote optimization of syllable structure

morphs that promote the information structure of words

16 reanalyses re-analyses follow syllabic principles (Swed. ni, lux. mir, dir nis)

reanalyses are not syllabically motivated (OHG ni.mis.du > ni.mist > NHD nimmst)

A. Dutch: a monotonic change towards a word language; differences between the North and the South

I. Evolution of Dutch towards a word language.

Dutch = Low Franconian with North Sea Germanic influences. The evolution is roughly parallel to that of High German:

i. Development of boundary signals

- Already in Old Dutch (OD): Final Devoicing (Quak 1997:43)

(1) wort (Wachtendonck Psalms (WP, around 1000 AD, 18.2) vs. gen. wordes ‘word’ gif, inf. geuon, (WP 59.13, 61.13) ‘to give’ weh (WP 18.6) (alongside weg) ‘way’

(This seems to confirm Vaught’s (1979) conclusion (reported in Szczepaniak 2007: 219-224) that there was final devoicing already in the OHG/OD period.)

- consonant epenthesis at word/morpheme endings (but more limited than in High German): ieman > iemant/d (criterion 12)

- development of linking s (criteria 12, 15) just as in High German

(2) a. raad#heer > raadsheer ‘lawyer’ b. dorp#café > dorpscafé ‘village pub’

ii. Degemination (criterion 4) in the early Middle Dutch (MD) (Van Bree 1987:155) (orthographic geminates however maintained in order to distinguish between tense and lax vowels). This paves the way for MD being a language with quantity based stress, analogous to the development toward MHG as mentioned in Szczepaniak (2007:124))

iii. Vowel Reduction (criterion 11) from the MD period onwards. Unlike Modern High

German, it is still productive. banaan > [bəˈnaːn]

iv. Syncope of unstressed vowels (criteria 1, 11): MD (hi) cussedese > (hi) cussese ‘he kissed her’

New Dutch (ND): poˈli[ts]ie > [pǝˈlisi] > [ˈplisi]. (Van Bree 1987:79)

Opposite typological developments in French and Dutch

- 3 -

v. Diphthongisation (criterion 9) of stressed tense high vowels in ND (15th-17th century)

(3) /y/: huus > huis , [hys] > [hœys] ‘house’ /i/: tijd [tit] > [tɛit] ‘time’ /u/: jou [ju] >[jɔu] ‘you’ II. Division between the North and the South: i. Apocope (criterion 11)

(4) MD sonne > ND zon ‘sun’, MD (ic) geve > ND (ik) geef ‘(I) give’ MD zone > ND zoon ‘son’

This change did not take place in all areas, Van Heule (1626, one of the first grammars of Dutch), notes that e-apocope happens in Hollandic Dutch (i.e. the Dutch spoken in the historic county of Holland), but not in Southern dialects like Flemish. This shows that in the early 17th century, this type of vowel deletion happened in Northern Dutch, but not in Southern Dutch:

“De Vlaemsche sprake die verlangt de woorden dikwils met eene E, zeggende Loopene, Draegene, Komene, Werkene, ook zeggen zy Schuyte, Wagene, Steene, zo ziet men hier dat de Hollantsche ende Vlaemsche sprake zeer strijdig is / de een de woorden te zeer verkortende / ende de ander die buyten natuyr verlangende.” (1626:91)

The Flemish language often demands word with an E, saying Loopene, Draegene, Komene, Werkene. They also say Draegene, Komene, Werkene. One thus sees here that the Hollandic and the Flemish language very much oppose each other, the one too much shortening and the other lengthening beyond what is natural. (my translation)

A systematic computer-aided analysis of notarial records, conducted by Van Reenen & Mulder (2003) show how apocope proliferated in the 14th century in Hollandish, while final ə remained intact in Flemish, cf. the following maps (Van Reenen & Mulder (2003:190-192, the lighter the area, the higher the percentage of apocope):

Roland Noske

- 4 -

Opposite typological developments in French and Dutch

- 5 -

Roland Noske

- 6 -

(map: the dialects of Middle Dutch)

Confirmation by the investigation of place names:

(5) Flanders / Zealand (rest of the) Netherlands Blankenberge HardenbergØ Oostende WesteindØ, OosterendØ, PurmerendØ Middelkerke OuwerkerkØ Oudenaarde DodewaardØ, TernaardØ Roeselare GelselaarØ Jabbeke JabeekØ IJzendijke GanzendijkØ Zuiddorpe, Westdorpe SlootdorpØ, RansdorpØ Zeebrugge BalkbrugØ Beervelde BentveldØ

ii. Diphthongisation of tense mid vowels in ‘Postmodern’ Northern Dutch (20th and 21st centuries (‘Poldernederlands’ (‘Polder Dutch’), Stroop 1998, Jacobi 2009):

/eː/ and /oː/: [eː] > [eːi]; [oː] > [oːu] (along with concomitant lowering [ɛi] > [ai], [ɔu] > [au], as part of a chain shift

(eː > eːi, ɛi > ai) to give way to the new Poldernederlands diphthongs). Not present in the Southern Dutch.

Opposite typological developments in French and Dutch

- 7 -

iii. Cliticization:

Word boundaries were blurred in the spelling of Middle Dutch: Van der Wal (1992:131) notes that there are many examples of proclitic and enclitic forms in Middle Dutch (MD) texts. This shows that word boundaries were often not felt:

(6) MD cliticized forms MD non-cliticized equivalents gloss a. tien tiden te dien tiden ‘at that time’ b. darme man die arme man ‘the poor man’

c. hi leidene hi leide ene ‘he lead him’ Multiple syllable induced vowel and consonant deletions, cliticizations, in Modern Colloquial Flemish, much more than in Northern Dutch;

(7) a. dat ik /dɑt ɩk/ > [dɑk b. was het /wɑs ət/ [wɑst] c. dat /dɑt/ > [dɑ] d. worden /wɔrd+ən/ => [wɔrən]

iv. Laison/glottal stop insertion

Contrast Modern Northern / Modern Southern Dutch: glottal stop insertion vs. liaison (Noske 2005:476):

(8) underlying form Northern Dutch Southern Dutch

a. /œyt+ɛində+lək/ [œyt.ˈʔɛində.lək] [œy.ˈtɛində.lək] b. /vər+ɑrm+ən/ [vər.ˈʔɑr.mən] [və.ˈrɑr.mən] c. /ɔn+eːns/ [ɔn.ˈʔeːns] [ɔ.ˈneːns] d. /bɛrɣ+ɑx.təx/ [ˈbɛrx.ˌʔɑx.təx] [ˈbɛr.ˌɣɑx.təx]

Conclusion: The facts of (i) apocope, (ii) diphthongisation of tense mid vowels, (iv) cliticization and (iv) liaison/glottal stop insertion suggest that the evolution towards a word language went further in the North than in the South. --------- B. French: a zig-zag evolution (Jacobs 1992:, ‘Pendular’ movement: of syllable structure typology: >closed (Class. Latin) > open (Late Latin) >closed (Gallo-Romance, Old French > much more open (Modern French)

Not only movement in syllable structure, but also in the word/syllable language typology:

Roland Noske

- 8 -

1. Early/Classical Latin: features showing clear characteristics of a word language:

i. The fusional type of language, with substantial freedom of word order substantial and the near-absence of clitics suggest that there was a high autonomy of the word. Little or no resyllabification. (criterion 13) (Marotta 1999:301, Sampson 2010:47).

ii. (regarding criterion 1) Complex maximal syllable structure: sCCVVCCs (Marotta 1999, Sampson 2010:46)

iii Vowel length contrast

iv. Syncope (criteria 1&11): 6th-5th cent.: in the 2nd syllable, caused by initial initial stress (Meiser 1998:53),

(9) *monestrum (related to moneō) > monstrum (Meiser 1998:66);

in the final syllable: after -t, mostly if the preceding syllable contains at least 2 moras:

(10) *dōt-is > dōs, *mort-is > mors (Meiser 1998:73)

v. Apocope (criteria 1&11):

(11) *h1éss(s)i > es, *h1éss-ti > est, *h1s-énti > sunt

(12) occasional apocope: fac, fer, dic, duc < face, *fere. dīce, dūce,

vi. Weakening (criterion 11): vowels in unstressed open syllables weakened to ə, later usually evolving to e (Meiser 1998:67)

ex.: internal syllables

(13) a >i: cad-ō ce-cid-ī b. e>i: leg-ō ē-lig-ō c. o>i *kupido-tāt-s > cupiditās (compare *kupid-os cupidus d. u>i: caput capit-is

+ 7 other types of vowel weakening in internal unstressed open syllables (Meiser 1998:68-70)

(14) in final syllables *pro-dāt > predit

vii. (criterion 6) Development towards quantity sensitive stress: from initial to penultimate/antepenultimate stress, dependent on the weight of the penultimate syllable.

However: already cluster simplification from Pre-Latin to Classical Latin (Maniet

1975:97-98, Steriade 1988, quoted by Sampson 2010:49, Maniet 1975)

(15) word-internally: a. *sarpmentum > sarmentum (cf. sarpi-ō) b. *indulgtum > indultum (cf. indulge-ō) c. *aksla > * azla > āla (cf. axilla) word-initally: d. *ktunika> tūnica ‘tunic’ e. *psaflom>sābulum ‘sand’ f. *sni > nix ’’snow’ word-finally g. *kord > cor

Opposite typological developments in French and Dutch

- 9 -

2. Classical Latin Late Latin: Clear movement into the direction of a syllable language

i Loss of vowel length distinction

ii. Monophtongization (already partially in Classical Latin, Väänänen 1984:38-39).

iii. Cluster simplification, giving rise to a more open syllable structure (Väänänen 1984:62-63):

(16) a. sanctus > santus, b. cinctus > cintus b. (dē)functus >defuntus c. sursus > susus, nunc > nuc d. quondam > quodam

iv. Loss of final nasals, giving rise to open syllable structure:

(17) diem > die (Väänänen 1984:66).

v . Beginnings of I-prosthesis (Sampson 2010:54ff), resolving sC(L) onsets.

(18) a. spes > ispes b. spatium >ispatium c. instutui > inistutui 3. Late Latin Gallo-Romance and Old French: A swing back to a word language

i. Second diphtongization (criterion 9): stressed vowels in open syllables are diphtongized (Fouché 1956, vol. 2:223ff, Pope 1953: 60-62, 103-104, La Chaussée 1989:182, 185, 187, 194).

(19) a. dēbet > OF deit (Mod. French doit,cf. devoir vs. doit) ‘must’ b. cor [kɔr] > [kuɔ̯r] ‘heart’ c. mare [ˈmaːre] > [ˈmaɛ̯rə] ‘sea’

ii. Degemination (criterion 4): The geminate consonants of Late Latin and those resulting from assimilation in the Gallo-Romance period were reduced to single consonants in Old French. This happened from the 9th century onwards (Pope 1952:147

iii. Vowel reduction due to stress (criterion 11, Pope 1952: 103-104)

(20) a. terːɑ > tɛrə ’earth’ b. portɑs > portəs ‘doors’

iv. Vowel deletion due to stress (criterion 11, Pope 1952: 112)

(21) a. perdere > pɛrdrə ‘loose’ b. ɑrborem > arbrə ‘tree’

v. Final devoicing (criterion 8) from the 7th century onwards (Pope 1952:98), centuries before the same process happened in Germanic ,

(22) a. [luŋɡum] > [luŋk] ‘long’ b. [ɡrɑndem] > [ɡrãnt] ‘big’ c. [perdo] > [pɛrt] ‘loose’ d. [riːsum] > *[rizʊ] > [ris] ‘laugh’ e. [serwum] > *[sɛrvə] > [sɛrf] ‘slave, servant’

4. Old French Middle French and Modern French: A swing to a very clear syllable language

i. Monophtongization (criterion 9) again, for the 11th century onwards (Pope 1952: 443ff), au >[o], ai >[e], etc.

Roland Noske

- 10 -

ii. Loss of productivity of vowel reduction (criterion 11): Unstressed full vowels introduced into the language at the time, e.g. in loan words, were no longer reduced to schwa. Schwas remain as relics of a word language

iii. Loss of final devoicing. (Apart from a few relics like the lexicalized alternations in neuf ~ neuve and ‘new’ grand ami [ɡrɑ̃tami] ~ grande amie [ɡrɑ̃dami] ‘big friend’.

iv. A change in the conditioning of final vowel deletion (criterion 11)

Vowel deletion (of schwa) has remained in Middle and Modern French, but has taken another role: instead of being conditioned by stress, it is now conditioned by syllable structure optimization:

(23) le + homme > l’homme (/lə+ɔm/ > [lɔm]) ‘the man, mankind’

proliferation of cliticization and resyllabification, v. Final Coronation of syllable-languagehood: loss of word accent. Because of vowel

reduction en syncope in OF, stress could become word final. From ± the 18 th century onwards: loss of word accent. The word ceases to be a relevant metric category. Development of Unbounded feet.

References: Auer, Peter, 1993. Is a rhythm-based typology possible? = KontRI Arbeitspapier Nr. 21,

Universität Konstanz. Downloadable from: http://www.germanistik.uni-freiburg.de/auer/ ?Personal :Prof._Dr._Peter_Auer:Publikationen#Online.

----. 2001. “Silben- und akzentzählende Sprachen”. In: Haspelmath, Martin, Ekkehard König, Wulf Oesterreicher & Wolfgang Raible (eds.). Language Typology and Language Universals. An International Handbook. Berlin: de Gruyter, 1391-1399.

Bree, Cor van. 1987. Historische Grammatica van het Nederlands. Dordrecht: Foris. On line available: http://www.dbnl.org/tekst/bree001hist02_01

Fouché, Pierre. 1958 Phonétique historique du français. 3 vols. Paris: Klincksiek. Heule, Christaen van. 1626. De Nederduytsche Spraec-konst ofte Taelbeschryvinghe. Leiden:

Jacob Roels. On line: http://www.dbnl.org/tekst/heul001nede01_01/. Facsimile edition of the 2nd edition (1633), edited and commented by W.J.H. Caron, published in 1953 by Wolters, Groningen.

Jacobi, Irene. 2009. On Variation and Change in Diphthongs and Long Vowels of Spoken Dutch. Ph.D. Dissertation, University of Amsterdam. On line available:

http://dare.uva.nl/record/293297 Jacobs, Haike. 1992. “The interaction between the evolution of syllable structure and foot

structure in the historical phonology of French”. In Terrel Morgan & Christiane Laueffer (eds.), Theoretical Analyses in Romance Linguistics. Amsterdam: John Benjamins, 55-79.

La Chaussée, F. de. 1989. Initation à la phonétique de l’ancien français. Nouvelle édition revue et augmentée. Paris : Klincksieck

Lahiri, Aditi, Thomas Riad & Haike Jacobs. 1999. ‘Chapter 6: Diachronic Prosody’. In: Van der Hulst, Harry.Word Prosodic Systems in the Languages of Europe. Berlin: De Gruyter, 335-422.

Opposite typological developments in French and Dutch

- 11 -

Meiser, Gerhard. 1998. Historische Laut- und Formenlehre der lateinischen Sprache. Darmstadt: Wissenschaftliche Buschgesellschaft.

Maniet, Albert. 1975. La phonétique historique du latin. 5ème éd. Paris: Klincksieck. Marotta, Giovanna. 1999. “The Latin Syllable”. In Van der Hulst, Harry & Nancy Ritter

(eds.), The Syllable. Views and Facts. Berlin: Mouton de Gruyter, 285-310. Noske, Roland. 2005. “A prosodic contrast between Northern and Southern Dutch: a result of

a Flemish-French sprachbund”. In Broekhuis, Hans et al., Organizing grammar. Linguistic Studies in Honor of Henk van Riemsdijk. Berlin: Mouton de Gruyter, 474-482. On line available: http://perso.univ-lille3.fr/~rnoske

----. 2009. “Autonomous typological prosodic evolution versus the Germanic superstrate in diachronic French phonology”. In Aboh, Enoch, Elisabeth van der Linden, Josep Quer & Petra Sleeman (eds.). Romance Languages and Linguistic Theory 2007, 223-242. Amsterdam/Philadelphia: Benjamins. On line available: http://perso.univ-lille3.fr/~rnoske

Nübling, Damaris & Renate Schrambke. 2004. “Silben- versus akzentsprachliche Züge in germanischen Sprachen und im Alemannischen”. In: Glaser, Elvira, Peter Ott & Rudolf Schwarzenbach (Hrsg.), Alemannisch im Sprachvergleich. Beiträge zur 14. Arbeitstagung für alemannische Dialektologie in Männedorf (Zürich) vom 16.-18.9.2002. Stuttgart: Franz Steiner, 280-320.

Quak, Arend. 1997. “2. Oudnederlands (tot circa 1200) door A. Quak met syntaxis door J.M. van der Horst”: ‘2.2. Interne taalgeschiedenis’. In: Toorn, Maarten C. van den et al. (red.), Geschiedenis van de Nederlandse taal. Amsterdam: Amsterdam University Press, 42-68. On line available: http://www.dbnl.org/tekst/toor004gesc01_01/

Pope, Mildred. 1952. From Latin to Modern French. 2nd edition. Manchester: Manchester University Press.

Reenen, Pieter van & Maaike Mulder. 2003. “Linguistic Interpretation of Spelling Variation and Spelling Conventions on the Basis of Charters in Middle Dutch and Old French: Methodological Aspects and Three Illustrations.” In: Goyens, Michèle & Werner Verbeke (eds.), The Dawn of the Written Vernacular in Western Europe. Leuven: Leuven University Press, 179-199.

Sampson, Rodney. 2010. Vowel Prosthesis in Romance. A Diachronic Study. Oxford: OUP. Steriade, Donca. 1988. “Gemination and the Proto-Romance Syllable Shift”. in Birdsong,

David and Jean-Pierre Montreuil (eds.). Advances in Romance Linguistics. Dordrecht: Foris.

Stroop, Jan. 1998. Poldernederlands: waardoor het ABN verdwijnt. Amsterdam: Bakker. Szczepaniak, Renata. 2007. Der phonologisch-typologische Wandel des Deutschen von einer

Silben- zu einer Wortsprache. Berlin: de Gruyter. Väänänen, Veikko. 1984. Introduction au Latin Vulgaire. 3ème éd. rev. et augm. Paris:

Klincksieck. Vaught, George M. 1979. A Study of Auslautverhärtung in Old High German. Ann Harbor. Wal, Marijke van der. 1992. Geschiedenis van het Nederlands. Utrecht: Het Spectrum.